HS project

Zoom in IBL and mORS cells

This file is used to analyse the immune cells dataset.

library(dplyr)
library(patchwork)
library(ggplot2)
library(ComplexHeatmap)
library(org.Mm.eg.db)

.libPaths()

## [1] "/usr/local/lib/R/library"

Preparation

In this section, we set the global settings of the analysis. We will store data there :

save_name = "iblmors"
out_dir = "."

We load the dataset :

sobj = readRDS(paste0(out_dir, "/", save_name, "_sobj.rds"))
sobj

## An object of class Seurat 
## 16701 features across 3532 samples within 1 assay 
## Active assay: RNA (16701 features, 2000 variable features)
##  6 dimensional reductions calculated: RNA_pca, RNA_pca_20_tsne, RNA_pca_20_umap, harmony, harmony_20_umap, harmony_20_tsne

We load the sample information :

sample_info = readRDS(paste0(out_dir, "/../../1_metadata/hs_hd_sample_info.rds"))
project_names_oi = sample_info$project_name

graphics::pie(rep(1, nrow(sample_info)),
              col = sample_info$color,
              labels = sample_info$project_name)

Here are custom colors for each cell type :

color_markers = readRDS(paste0(out_dir, "/../../1_metadata/hs_hd_color_markers.rds"))

data.frame(cell_type = names(color_markers),
           color = unlist(color_markers)) %>%
  ggplot2::ggplot(., aes(x = cell_type, y = 0, fill = cell_type)) +
  ggplot2::geom_point(pch = 21, size = 5) +
  ggplot2::scale_fill_manual(values = unlist(color_markers), breaks = names(color_markers)) +
  ggplot2::theme_classic() +
  ggplot2::theme(legend.position = "none",
                 axis.line = element_blank(),
                 axis.title = element_blank(),
                 axis.ticks = element_blank(),
                 axis.text.y = element_blank(),
                 axis.text.x = element_text(angle = 30, hjust = 1))

This is the projection of interest :

name2D = "harmony_20_tsne"

We design a custom function to make the GSEA plot and a word cloud graph :

make_gsea_plot = function(gsea_results, gs_oi, fold_change, metric = "FC") {
  fold_change$metric = fold_change[, metric]
  
  plot_list = lapply(gs_oi, FUN = function(gene_set) {
    # Gene set content
    gs_content = gene_sets %>%
      dplyr::filter(gs_name == gene_set) %>%
      dplyr::pull(ensembl_gene) %>%
      unique()
    
    # Gene set size
    nb_genes = length(gs_content)
    
    # Enrichment metrics
    NES = gsea_results@result[gene_set, "NES"]
    p.adjust = gsea_results@result[gene_set, "p.adjust"] %>%
      round(., 4)
    qvalues = gsea_results@result[gene_set, "qvalues"]
    
    if (p.adjust > 0.05) {
      p.adjust = paste0("<span style='color:red;'>", p.adjust, "</span>")
    }
    
    my_subtitle = paste0("\nNES : ", round(NES, 2),
                         " | padj : ", p.adjust,
                         " | qval : ", round(qvalues, 4),
                         " | set size : ", nb_genes, " genes")
    
    # Size limits
    lower_FC = min(fold_change[gs_content, ]$metric, na.rm = TRUE)
    upper_FC = max(fold_change[gs_content, ]$metric, na.rm = TRUE)
    
    # Plot
    p = enrichplot::gseaplot2(x = gsea_results, geneSetID = gene_set) +
      ggplot2::labs(title = gene_set,
                    subtitle = my_subtitle) +
      ggplot2::theme(plot.title = element_text(hjust = 0.5, face = "bold",
                                               margin = ggplot2::margin(3, 3, 5, 3)),
                     plot.subtitle = ggtext::element_markdown(hjust = 0.5,
                                                              size = 10))
    
    wc = ggplot2::ggplot(fold_change[gs_content, ],
                         aes(label = gene_name, size = abs(metric), color = metric)) +
      ggwordcloud::geom_text_wordcloud_area(show.legend = TRUE) +
      ggplot2::scale_color_gradient2(
        name = metric,
        low = aquarius::color_cnv[1],
        mid = "gray70", midpoint = 0,
        high = aquarius::color_cnv[3]) +
      ggplot2::scale_size_area(max_size = 7) +
      ggplot2::theme_minimal() +
      ggplot2::guides(size = "none")
    
    return(list(p, wc))
  }) %>% unlist(., recursive = FALSE)
  
  return(plot_list)
}

Visualization

Gene expression

We visualize gene expression for some markers :

features = c("percent.mt", "percent.rb", "nFeature_RNA")

plot_list = lapply(features, FUN = function(one_gene) {
  Seurat::FeaturePlot(sobj, features = one_gene,
                      reduction = name2D) +
    ggplot2::theme(aspect.ratio = 1) +
    ggplot2::scale_color_gradientn(colors = aquarius::color_gene) +
    Seurat::NoAxes()
})

patchwork::wrap_plots(plot_list, ncol = 3)

Clusters

We visualize clusters :

cluster_plot = Seurat::DimPlot(sobj, reduction = name2D, label = TRUE) +
  Seurat::NoAxes() +
  ggplot2::theme(aspect.ratio = 1)
cluster_plot

Cell type

We visualize cell type split by sample :

plot_list = aquarius::plot_split_dimred(sobj,
                                        reduction = name2D,
                                        split_by = "project_name",
                                        group_by = "cell_type",
                                        split_color = setNames(sample_info$color,
                                                               nm = sample_info$project_name),
                                        group_color = color_markers,
                                        bg_pt_size = 0.5, main_pt_size = 0.5)

plot_list[[length(plot_list) + 1]] = cluster_plot

patchwork::wrap_plots(plot_list, ncol = 4) &
  Seurat::NoLegend()

Cluster type

We summarize major cell type by cluster :

cell_type_clusters = sobj@meta.data[, c("cell_type", "seurat_clusters")] %>%
  table() %>%
  prop.table(., margin = 2) %>%
  apply(., 2, which.max)
cell_type_clusters = setNames(levels(sobj$cell_type)[cell_type_clusters],
                              nm = names(cell_type_clusters))

We define cluster type :

sobj$cluster_type = cell_type_clusters[sobj$seurat_clusters] %>%
  as.factor()
table(sobj$cluster_type, sobj$cell_type)

##      
##       CD4 T cells CD8 T cells Langerhans cells macrophages B cells cuticle
##   IBL           0           0                0           0       1       1
##   ORS           2           0                0           0       3       7
##      
##       cortex medulla  IRS proliferative  IBL  ORS  IFE HFSC sebocytes
##   IBL      0       0    0             9 1465    5    7   19         3
##   ORS      2      20    7            37   26 1785   65   48        20

We subset color_markers :

color_markers = color_markers[levels(sobj$cluster_type)]

We compare cluster annotation and cell type annotation :

p1 = Seurat::DimPlot(sobj, group.by = "cell_type",
                     reduction = name2D, cols = color_markers) +
  ggplot2::labs(title = "Cell type") +
  Seurat::NoAxes() +
  ggplot2::theme(aspect.ratio = 1,
                 plot.title = element_text(hjust = 0.5))

p2 = Seurat::DimPlot(sobj, group.by = "cluster_type",
                     reduction = name2D, cols = color_markers) +
  ggplot2::labs(title = "Cluster type") +
  Seurat::NoAxes() +
  ggplot2::theme(aspect.ratio = 1,
                 plot.title = element_text(hjust = 0.5))

patchwork::wrap_plots(p1, p2, guides = "collect")

Barplot

We make a barplot to compare IBL and ORS populations in HS vs HD samples. The proportion of IBL is indicated on top of bars.

quantif = table(sobj$sample_identifier,
                sobj$cluster_type) %>%
  as.data.frame.table() %>%
  `colnames<-`(c("Sample", "cell_type", "nb_cells")) %>%
  dplyr::group_by(Sample) %>%
  dplyr::mutate(total_cells = sum(nb_cells)) %>%
  as.data.frame() %>%
  dplyr::filter(cell_type == "IBL") %>%
  dplyr::mutate(prop_ibl = nb_cells / total_cells) %>%
  dplyr::mutate(prop_ibl = 100*round(prop_ibl, 4))

sobj$seurat_clusters = factor(sobj$seurat_clusters,
                              levels = names(sort(cell_type_clusters)))

aquarius::plot_barplot(df = table(sobj$sample_identifier,
                                  sobj$seurat_clusters) %>%
                         as.data.frame.table() %>%
                         `colnames<-`(c("sample_identifier", "clusters", "nb_cells")),
                       x = "sample_identifier", y = "nb_cells", fill = "clusters",
                       position = position_fill()) +
  ggplot2::scale_fill_manual(values = c(colorRampPalette(c("chartreuse1", "chartreuse4"))(table(sort(cell_type_clusters))["IBL"]),
                                        colorRampPalette(c("cadetblue1", "cadetblue4"))(table(sort(cell_type_clusters))["ORS"])),
                             breaks = names(sort(cell_type_clusters)),
                             name = "Cell type") +
  ggplot2::geom_label(data = quantif, inherit.aes = FALSE,
                      aes(x = .data$Sample, y = 1.05, label = .data$prop_ibl),
                      label.size = 0, size = 5)

Differential expression

In this section, we perform DE between :

• inner bulge layer (IBL) and outer root sheath (ORS) populations
• healthy donors (HD) and HS patients (HS) within ORS population
• healthy donors (HD) and HS patients (HS) within IBL population
• cluster 5 vs rest of ORS, because this cluster is specific to HS compared to HD

We save the results in a list :

list_results = list()

We make over-representation analysis for each group of genes. We load gene sets from MSigDB :

gene_sets = aquarius::get_gene_sets(species = "Homo sapiens")
gene_sets = gene_sets$gene_sets

head(gene_sets)

## # A tibble: 6 x 16
##   gs_cat gs_subcat gs_name gene_symbol entrez_gene ensembl_gene human_gene_symb~
##   <chr>  <chr>     <chr>   <chr>             <int> <chr>        <chr>           
## 1 C5     GO:BP     GOBP_1~ AASDHPPT          60496 ENSG0000014~ AASDHPPT        
## 2 C5     GO:BP     GOBP_1~ ALDH1L1           10840 ENSG0000014~ ALDH1L1         
## 3 C5     GO:BP     GOBP_1~ ALDH1L2          160428 ENSG0000013~ ALDH1L2         
## 4 C5     GO:BP     GOBP_1~ MTHFD1             4522 ENSG0000010~ MTHFD1          
## 5 C5     GO:BP     GOBP_1~ MTHFD1L           25902 ENSG0000012~ MTHFD1L         
## 6 C5     GO:BP     GOBP_1~ MTHFD2L          441024 ENSG0000016~ MTHFD2L         
## # ... with 9 more variables: human_entrez_gene <int>, human_ensembl_gene <chr>,
## #   gs_id <chr>, gs_pmid <chr>, gs_geoid <chr>, gs_exact_source <chr>,
## #   gs_url <chr>, gs_description <chr>, category <chr>

How many gene sets ?

gene_sets[, c("gs_subcat", "gs_name")] %>%
  dplyr::distinct() %>%
  dplyr::pull(gs_subcat) %>%
  table() %>%
  as.data.frame.table() %>%
  `colnames<-`(c("Category", "Nb gene sets"))

##          Category Nb gene sets
## 1                           50
## 2         CP:KEGG          186
## 3          CP:PID          196
## 4     CP:REACTOME         1615
## 5 CP:WIKIPATHWAYS          664
## 6           GO:BP         7658
## 7           GO:CC         1006
## 8           GO:MF         1738

We get gene name and gene ID correspondence :

gene_corresp = sobj@assays[["RNA"]]@meta.features[, c("gene_name", "Ensembl_ID")] %>%
  `colnames<-`(c("NAME", "ID")) %>%
  dplyr::mutate(ID = as.character(ID))
rownames(gene_corresp) = gene_corresp$ID

head(gene_corresp)

##                       NAME              ID
## ENSG00000238009 AL627309.1 ENSG00000238009
## ENSG00000237491 AL669831.5 ENSG00000237491
## ENSG00000225880  LINC00115 ENSG00000225880
## ENSG00000230368     FAM41C ENSG00000230368
## ENSG00000230699 AL645608.3 ENSG00000230699
## ENSG00000187634     SAMD11 ENSG00000187634

IBL vs ORS

group_name = "IBL_vs_ORS"

We change cell identities to cluster type :

Seurat::Idents(sobj) = sobj$cluster_type

table(Seurat::Idents(sobj), sobj$sample_type)

##      
##         HS   HD
##   IBL 1284  226
##   ORS 1563  459

DE

We identify specific markers for each population :

mark = Seurat::FindMarkers(sobj, ident.1 = "IBL", ident.2 = "ORS")

mark = mark %>%
  dplyr::filter(p_val_adj < 0.05) %>%
  dplyr::arrange(-avg_logFC, pct.1 - pct.2)

list_results[[group_name]]$mark = mark

dim(mark)

## [1] 1027    5

head(mark, n = 20)

##                  p_val avg_logFC pct.1 pct.2     p_val_adj
## KRT16     0.000000e+00  4.323925 0.920 0.116  0.000000e+00
## KRT6B     0.000000e+00  3.627468 0.921 0.285  0.000000e+00
## FABP5     0.000000e+00  3.348761 0.993 0.509  0.000000e+00
## KRT17     0.000000e+00  3.020736 0.987 0.733  0.000000e+00
## KRT6A    1.267920e-155  2.994730 0.764 0.476 2.117554e-151
## CST6     1.834243e-160  2.741295 0.391 0.038 3.063369e-156
## KRT6C    9.379153e-283  2.653075 0.568 0.038 1.566412e-278
## TMSB4X    0.000000e+00  2.356496 0.989 0.864  0.000000e+00
## LGALS1    0.000000e+00  2.320083 0.929 0.484  0.000000e+00
## S100A2    0.000000e+00  2.134213 0.996 0.977  0.000000e+00
## CALML3    0.000000e+00  2.003586 0.972 0.686  0.000000e+00
## GJA1      0.000000e+00  1.973282 0.875 0.586  0.000000e+00
## GJB6      0.000000e+00  1.907928 0.916 0.656  0.000000e+00
## SBSN     9.451796e-242  1.751399 0.534 0.047 1.578544e-237
## APOE      0.000000e+00  1.744652 0.983 0.772  0.000000e+00
## NDUFA4L2 1.058296e-239  1.730465 0.697 0.217 1.767461e-235
## LYPD3     0.000000e+00  1.680352 0.738 0.135  0.000000e+00
## TUBB2A    0.000000e+00  1.667800 0.756 0.181  0.000000e+00
## SDC1      0.000000e+00  1.662321 0.838 0.517  0.000000e+00
## GJB2     1.288894e-254  1.634612 0.826 0.595 2.152582e-250

There are 1027 genes differentially expressed. How many for each population ?

# IBL
mark_ibl = mark %>%
  dplyr::filter(avg_logFC > 0)
nrow(mark_ibl)

## [1] 434

# ORS
mark_ors = mark %>%
  dplyr::filter(avg_logFC < 0)
nrow(mark_ors)

## [1] 593

We represent the information on a figure :

mark$gene_name = rownames(mark)
mark_to_label = rbind(
  # up-regulated in IBL
  mark %>% dplyr::top_n(., n = 20, wt = avg_logFC),
  # up-regulated in ORS
  mark %>% dplyr::top_n(., n = 20, wt = -avg_logFC),
  # representative and selective for IBL
  mark %>% dplyr::top_n(., n = 20, wt = (pct.1 - pct.2)),
  # representative and selective for ORS
  mark %>% dplyr::top_n(., n = 20, wt = -(pct.1 - pct.2))) %>%
  dplyr::distinct()

ggplot2::ggplot(mark, aes(x = pct.1, y = pct.2, col = avg_logFC)) +
  ggplot2::geom_abline(slope = 1, intercept = 0, lty = 2) +
  ggplot2::geom_point() +
  ggrepel::geom_label_repel(data = mark_to_label, max.overlaps = Inf,
                            aes(x = pct.1, y = pct.2, label = gene_name),
                            col = "black", fill = NA, size = 3, label.size = NA) +
  ggplot2::labs(title = "Differentially expressed genes",
                subtitle = "between IBL and ORS") +
  ggplot2::scale_color_gradient2(low = aquarius::color_cnv[1],
                                 mid = aquarius::color_cnv[2],
                                 high = aquarius::color_cnv[3],
                                 midpoint = 0) +
  ggplot2::theme_classic() +
  ggplot2::theme(plot.title = element_text(hjust = 0.5),
                 plot.subtitle = element_text(hjust = 0.5))

We represent the best genes on a violin plot, group by cluster type and split by sample type :

mark_to_label = mark_to_label %>%
  dplyr::arrange(pct.1 - pct.2)

plot_list = lapply(mark_to_label$gene_name, FUN = function(gene) {
  p = Seurat::VlnPlot(sobj, features = gene, pt.size = 0.001,
                      group.by = "cluster_type",
                      split.by = "sample_type") +
    ggplot2::scale_fill_manual(breaks = c("HS", "HD"),
                               values = c("#C55F40", "#2C78E6")) +
    ggplot2::theme(axis.title.x = element_blank(),
                   axis.title.y = element_blank(),
                   legend.position = "none")
  return(p)
})

patchwork::wrap_plots(plot_list, ncol = 5)

On the figure, we can see that some specific markers for a population are indeed specific to a sample type rather than the whole population.

IBL

We explore enrichment in gene sets for IBL population

genes_of_interest = rownames(mark_ibl)

enrichr_results = aquarius::run_enrichr(gene_names = genes_of_interest,
                                        gene_corresp = gene_corresp,
                                        gene_sets = gene_sets[, c("gs_name", "ensembl_gene")],
                                        make_plot = TRUE,
                                        plot_title = "Up-regulated in IBL compared to ORS")

list_results[[group_name]]$enrichr_ibl = enrichr_results$ego

enrichr_results$plot +
  ggplot2::theme(axis.text.y = element_text(size = 8))

ORS

We explore enrichment in gene sets for ORS population.

genes_of_interest = rownames(mark_ors)

enrichr_results = aquarius::run_enrichr(gene_names = genes_of_interest,
                                        gene_corresp = gene_corresp,
                                        gene_sets = gene_sets[, c("gs_name", "ensembl_gene")],
                                        make_plot = TRUE,
                                        plot_title = "Up-regulated in ORS compared to IBL")

list_results[[group_name]]$enrichr_ors = enrichr_results$ego

enrichr_results$plot +
  ggplot2::theme(axis.text.y = element_text(size = 8))

GSEA

We run a GSEA for all gene sets, from the full count matrix :

ranked_gene_list = aquarius::run_foldchange(Seurat::GetAssayData(sobj, assay = "RNA", slot = "counts"),
                                            group1 = colnames(sobj)[sobj@active.ident %in% "IBL"],
                                            group2 = colnames(sobj)[sobj@active.ident %in% "ORS"])
names(ranked_gene_list) = gene_corresp$ID

gsea_results = aquarius::gsea_run(ranked_gene_list = ranked_gene_list,
                                  gene_sets = gene_sets[, c("gs_name", "ensembl_gene")],
                                  GSEA_p_val_thresh = 1)

list_results[[group_name]]$gsea = gsea_results

gsea_results@result %>%
  dplyr::filter(pvalue < 0.05) %>%
  dplyr::top_n(., n = 200, wt = abs(NES)) %>%
  dplyr::mutate(too_long = ifelse(nchar(ID) > 60, yes = TRUE, no = FALSE)) %>%
  dplyr::mutate(ID = stringr::str_sub(ID, end = 60)) %>%
  dplyr::mutate(ID = ifelse(too_long, yes = paste0(ID, "..."), no = ID)) %>%
  aquarius::gsea_plot(show_legend = TRUE) +
  ggplot2::labs(title = "GSEA using all genes (count matrix)") +
  ggplot2::theme(plot.title = element_text(size = 20))

We make the gsea plot for two gene sets :

p1 = enrichplot::gseaplot2(x = gsea_results, geneSetID = "REACTOME_KERATINIZATION") +
  ggplot2::labs(title = "REACTOME_KERATINIZATION") +
  ggplot2::theme(plot.title = element_text(hjust = 0.5, face = "bold",
                                           margin = ggplot2::margin(3, 3, 5, 3)))

p2 = enrichplot::gseaplot2(x = gsea_results, geneSetID = "HALLMARK_INTERFERON_GAMMA_RESPONSE") +
  ggplot2::labs(title = "HALLMARK_INTERFERON_GAMMA_RESPONSE") +
  ggplot2::theme(plot.title = element_text(hjust = 0.5, face = "bold",
                                           margin = ggplot2::margin(3, 3, 5, 3)))

p1 | p2

Maybe those gene sets are specific to sample type rather than the whole cell population.

Cluster 5 within ORS

group_name = "cluster5_vs_ORS"

We subset the Seurat object and change cell identities to sample type.

subsobj = subset(sobj, cluster_type == "ORS")
subsobj$seurat_clusters = base::droplevels(subsobj$seurat_clusters)
Seurat::Idents(subsobj) = subsobj$seurat_clusters

table(subsobj$seurat_clusters)

## 
##   0   1   3   5   7 
## 669 438 384 302 229

DE

We identify specific markers for each population :

mark = Seurat::FindMarkers(subsobj, ident.1 = 5)

mark = mark %>%
  dplyr::filter(p_val_adj < 0.05) %>%
  dplyr::arrange(-avg_logFC, pct.1 - pct.2)

list_results[[group_name]]$mark = mark

dim(mark)

## [1] 395   5

head(mark, n = 20)

##                  p_val avg_logFC pct.1 pct.2     p_val_adj
## SLPI     7.619364e-128 1.5306935 0.765 0.164 1.272510e-123
## MT1X      4.989262e-73 1.4231367 0.954 0.680  8.332567e-69
## KRT15    2.275685e-123 1.3705217 1.000 0.874 3.800622e-119
## LY6D      1.984327e-88 1.2250993 0.503 0.080  3.314025e-84
## IGFBP3   1.494292e-140 1.1855371 0.437 0.016 2.495618e-136
## IFI27     1.438175e-99 1.1442636 0.556 0.085  2.401895e-95
## MT1E      2.040603e-70 1.0839955 0.957 0.659  3.408012e-66
## WNT3     1.095829e-212 1.0819468 0.924 0.134 1.830143e-208
## AQP3      5.658795e-83 1.0173107 0.977 0.733  9.450753e-79
## NEAT1     3.110435e-52 0.9735049 0.990 0.905  5.194737e-48
## TSC22D3   1.298633e-75 0.9601088 0.742 0.258  2.168848e-71
## MTRNR2L1  4.400300e-32 0.9329154 0.815 0.593  7.348942e-28
## CXCL14    7.450833e-63 0.9282610 0.997 0.840  1.244364e-58
## ZFP36     1.579077e-42 0.9002595 0.742 0.406  2.637216e-38
## GLUL      7.969325e-58 0.8930699 0.934 0.699  1.330957e-53
## ID1       2.365404e-17 0.8903797 0.652 0.442  3.950461e-13
## FOS       1.096048e-36 0.8789679 0.967 0.837  1.830511e-32
## IL1R2    5.014612e-122 0.8759829 0.675 0.110 8.374904e-118
## IL18     6.622739e-101 0.8511305 0.970 0.456  1.106064e-96
## HOPX      1.939070e-75 0.8395659 0.871 0.417  3.238441e-71

There are 395 genes differentially expressed. How many for each population ?

# Cluster 5
mark_cluster5 = mark %>%
  dplyr::filter(avg_logFC > 0)
nrow(mark_cluster5)

## [1] 225

# Other ORS
mark_not5 = mark %>%
  dplyr::filter(avg_logFC < 0)
nrow(mark_not5)

## [1] 170

We represent the information on a figure :

mark$gene_name = rownames(mark)
mark_signif = rbind(
  # up-regulated in cluster 5
  mark %>% dplyr::top_n(., n = 20, wt = avg_logFC),
  # up-regulated in other ORS
  mark %>% dplyr::top_n(., n = 20, wt = -avg_logFC),
  # representative and selective for cluster 5
  mark %>% dplyr::top_n(., n = 20, wt = (pct.1 - pct.2)),
  # representative and selective for other ORS
  mark %>% dplyr::top_n(., n = 20, wt = -(pct.1 - pct.2))) %>%
  dplyr::distinct()

ggplot2::ggplot(mark, aes(x = pct.1, y = pct.2, col = avg_logFC)) +
  ggplot2::geom_abline(slope = 1, intercept = 0, lty = 2) +
  ggplot2::geom_point() +
  ggrepel::geom_label_repel(data = mark_signif, max.overlaps = Inf,
                            aes(x = pct.1, y = pct.2, label = gene_name),
                            col = "black", fill = NA, size = 3, label.size = NA) +
  ggplot2::labs(title = "Differentially expressed genes",
                subtitle = "between cluster 5 and other ORS") +
  ggplot2::scale_color_gradient2(low = aquarius::color_cnv[1],
                                 mid = aquarius::color_cnv[2],
                                 high = aquarius::color_cnv[3],
                                 midpoint = 0) +
  ggplot2::theme_classic() +
  ggplot2::theme(plot.title = element_text(hjust = 0.5),
                 plot.subtitle = element_text(hjust = 0.5))

Heatmap

We represent a subset of genes on a heatmap :

features_oi = mark %>%
  dplyr::filter(p_val_adj < 0.05) %>%
  dplyr::filter(abs(avg_logFC) > 0.5) %>%
  rownames()

length(features_oi)

## [1] 112

We prepare the scaled expression matrix :

mat_expression = Seurat::GetAssayData(subsobj, assay = "RNA", slot = "data")[features_oi, ]
mat_expression = Matrix::t(mat_expression)
mat_expression = dynutils::scale_quantile(mat_expression) # between 0 and 1
mat_expression = Matrix::t(mat_expression)
mat_expression = as.matrix(mat_expression) # not sparse

dim(mat_expression)

## [1]  112 2022

We prepare the heatmap annotation :

ha_top = ComplexHeatmap::HeatmapAnnotation(
  cluster_type = subsobj$cluster_type,
  sample_type = subsobj$sample_type,
  cluster = subsobj$seurat_clusters,
  col = list(cluster_type = color_markers,
             sample_type = setNames(nm = c("HS", "HD"),
                                    c("#C55F40", "#2C78E6")),
             cluster = setNames(nm = levels(subsobj$seurat_clusters),
                                aquarius::gg_color_hue(length(levels(subsobj$seurat_clusters))))))

And the heatmap :

ht = ComplexHeatmap::Heatmap(mat_expression,
                             col = aquarius::color_cnv,
                             # Annotation
                             top_annotation = ha_top,
                             # Grouping
                             column_title = NULL,
                             cluster_rows = TRUE,
                             cluster_columns = TRUE,
                             show_column_names = FALSE,
                             # Visual aspect
                             show_heatmap_legend = TRUE,
                             border = TRUE)

ComplexHeatmap::draw(ht,
                     merge_legend = TRUE,
                     heatmap_legend_side = "bottom",
                     annotation_legend_side = "bottom")

Up-regulated in cluster 5

We explore enrichment in gene sets for cluster 5.

genes_of_interest = rownames(mark_cluster5)

enrichr_results = aquarius::run_enrichr(gene_names = genes_of_interest,
                                        gene_corresp = gene_corresp,
                                        gene_sets = gene_sets[, c("gs_name", "ensembl_gene")],
                                        make_plot = TRUE,
                                        plot_title = "Up-regulated in cluster 5 compared to other ORS")

list_results[[group_name]]$enrichr_up = enrichr_results$ego

enrichr_results$plot +
  ggplot2::theme(axis.text.y = element_text(size = 8))

Down-regulated in cluster 5

We explore enrichment in gene sets for genes downregulated in cluster 5.

genes_of_interest = rownames(mark_not5)

enrichr_results = aquarius::run_enrichr(gene_names = genes_of_interest,
                                        gene_corresp = gene_corresp,
                                        gene_sets = gene_sets[, c("gs_name", "ensembl_gene")],
                                        make_plot = TRUE,
                                        plot_title = "Down-regulated in cluster 5 compared to other ORS")

list_results[[group_name]]$enrichr_dn = enrichr_results$ego

enrichr_results$plot +
  ggplot2::theme(axis.text.y = element_text(size = 8))

GSEA

We run a GSEA for all gene sets, from the full count matrix :

ranked_gene_list = aquarius::run_foldchange(Seurat::GetAssayData(subsobj, assay = "RNA", slot = "counts"),
                                            group1 = colnames(subsobj)[subsobj@active.ident == 5],
                                            group2 = colnames(subsobj)[subsobj@active.ident != 5])
names(ranked_gene_list) = gene_corresp$ID

gsea_results = aquarius::gsea_run(ranked_gene_list = ranked_gene_list,
                                  gene_sets = gene_sets[, c("gs_name", "ensembl_gene")],
                                  GSEA_p_val_thresh = 1)

list_results[[group_name]]$gsea = gsea_results

gsea_results@result %>%
  dplyr::filter(pvalue < 0.05) %>%
  dplyr::top_n(., n = 200, wt = abs(NES)) %>%
  dplyr::mutate(too_long = ifelse(nchar(ID) > 60, yes = TRUE, no = FALSE)) %>%
  dplyr::mutate(ID = stringr::str_sub(ID, end = 60)) %>%
  dplyr::mutate(ID = ifelse(too_long, yes = paste0(ID, "..."), no = ID)) %>%
  aquarius::gsea_plot(show_legend = TRUE) +
  ggplot2::labs(title = "GSEA using all genes (count matrix)") +
  ggplot2::theme(plot.title = element_text(size = 20))

We compute a fold change table to make a wordcloud :

fold_change = gene_corresp
colnames(fold_change) = c("gene_name", "ID")
rownames(fold_change) = fold_change$ID
fold_change$FC = ranked_gene_list[rownames(fold_change)]
fold_change$pct.1 = Seurat::GetAssayData(subsobj, assay = "RNA", slot = "counts")[, subsobj@active.ident == "5"] %>%
  apply(., MARGIN = 1, FUN = function(one_row) {
    return( mean(one_row != 0) )
  })
fold_change$pct.2 = Seurat::GetAssayData(subsobj, assay = "RNA", slot = "counts")[, subsobj@active.ident != "5"] %>%
  apply(., MARGIN = 1, FUN = function(one_row) {
    return( mean(one_row != 0) )
  })
fold_change$FC_x_pct = ifelse(fold_change$FC > 0,
                              yes = fold_change$FC * fold_change$pct.1,
                              no = fold_change$FC * fold_change$pct.2)

dim(fold_change) ; head(fold_change)

## [1] 16701     6

##                  gene_name              ID           FC       pct.1       pct.2
## ENSG00000238009 AL627309.1 ENSG00000238009  0.998441346 0.009933775 0.004069767
## ENSG00000237491 AL669831.5 ENSG00000237491  0.985617306 0.162251656 0.055232558
## ENSG00000225880  LINC00115 ENSG00000225880  0.285723298 0.049668874 0.033720930
## ENSG00000230368     FAM41C ENSG00000230368 -1.335459391 0.036423841 0.065116279
## ENSG00000230699 AL645608.3 ENSG00000230699 -0.001558654 0.006622517 0.006395349
## ENSG00000187634     SAMD11 ENSG00000187634 -1.460990273 0.000000000 0.005232558
##                      FC_x_pct
## ENSG00000238009  9.918292e-03
## ENSG00000237491  1.599180e-01
## ENSG00000225880  1.419155e-02
## ENSG00000230368 -8.696015e-02
## ENSG00000230699 -9.968137e-06
## ENSG00000187634 -7.644717e-03

We make the gsea plot for some gene sets :

gs_oi = c("GOBP_CD4_POSITIVE_ALPHA_BETA_T_CELL_DIFFERENTIATION",
          "GOBP_REGULATION_OF_ALPHA_BETA_T_CELL_DIFFERENTIATION",
          "GOBP_ALPHA_BETA_T_CELL_DIFFERENTIATION",
          "GOBP_POSITIVE_REGULATION_OF_ALPHA_BETA_T_CELL_ACTIVATION",
          "GOBP_ALPHA_BETA_T_CELL_ACTIVATION")

plot_list = make_gsea_plot(gsea_results, gs_oi, fold_change, "FC_x_pct")

patchwork::wrap_plots(plot_list, ncol = 2, widths = c(2,1.5))

IBL : HS vs HD

group_name = "IBL_HS_vs_HD"

We subset the Seurat object and change cell identities to sample type.

subsobj = subset(sobj, cluster_type == "IBL")
Seurat::Idents(subsobj) = subsobj$sample_type

table(subsobj$sample_type)

## 
##   HS   HD 
## 1284  226

DE

We identify DE genes between HS and HD :

mark = Seurat::FindMarkers(subsobj, ident.1 = "HS", ident.2 = "HD")

mark = mark %>%
  dplyr::filter(p_val_adj < 0.05) %>%
  dplyr::arrange(-avg_logFC, pct.1 - pct.2)

list_results[[group_name]]$mark = mark

dim(mark)

## [1] 107   5

head(mark, n = 20)

##                  p_val avg_logFC pct.1 pct.2    p_val_adj
## LGALS7    1.216786e-49 1.6340915 0.685 0.190 2.032154e-45
## S100A7    2.913857e-11 1.5790394 0.289 0.080 4.866433e-07
## MIF       1.195470e-44 1.2201587 0.706 0.261 1.996554e-40
## LGALS7B   1.099708e-09 1.1642854 0.769 0.717 1.836623e-05
## MTRNR2L12 9.867130e-12 1.0580493 0.393 0.181 1.647909e-07
## MTRNR2L8  3.047443e-11 0.9124631 0.350 0.155 5.089535e-07
## MT-CO1    4.350400e-12 0.9105978 0.449 0.226 7.265603e-08
## FOSB      7.804218e-09 0.8760606 0.298 0.133 1.303382e-04
## NDUFA4L2  1.283179e-20 0.8201600 0.743 0.434 2.143037e-16
## MT-CO2    1.272279e-08 0.8121340 0.433 0.265 2.124832e-04
## S100A9    1.727393e-13 0.7487053 0.315 0.075 2.884919e-09
## KRT15     2.873193e-08 0.6857105 0.290 0.115 4.798519e-04
## CA2       1.116064e-16 0.6640743 0.460 0.173 1.863938e-12
## ARF5      3.584966e-27 0.6221800 0.433 0.044 5.987251e-23
## CKB       4.051432e-08 0.5928841 0.282 0.115 6.766297e-04
## MT-ND4    2.524958e-07 0.5633098 0.410 0.248 4.216932e-03
## MT-ATP6   1.537778e-06 0.5539256 0.496 0.358 2.568244e-02
## CST6      1.707459e-07 0.5234863 0.421 0.221 2.851627e-03
## IGKC      5.879378e-07 0.5045779 0.101 0.000 9.819150e-03
## MCL1      1.177137e-07 0.5038666 0.379 0.208 1.965937e-03

Up in HS

We explore enrichment in gene sets for HS population.

genes_of_interest = mark %>%
  dplyr::filter(avg_logFC > 0) %>%
  rownames()

enrichr_results = aquarius::run_enrichr(gene_names = genes_of_interest,
                                        gene_corresp = gene_corresp,
                                        gene_sets = gene_sets[, c("gs_name", "ensembl_gene")],
                                        make_plot = TRUE,
                                        plot_title = "Up-regulated in HS compared to HD")

list_results[[group_name]]$enrichr_hs = enrichr_results$ego

enrichr_results$plot +
  ggplot2::theme(axis.text.y = element_text(size = 8))

Up in HD

We explore enrichment in gene sets for HD population.

genes_of_interest = mark %>%
  dplyr::filter(avg_logFC < 0) %>%
  rownames()

enrichr_results = aquarius::run_enrichr(gene_names = genes_of_interest,
                                        gene_corresp = gene_corresp,
                                        gene_sets = gene_sets[, c("gs_name", "ensembl_gene")],
                                        make_plot = TRUE,
                                        plot_title = "Down-regulated in HS compared to HD")

list_results[[group_name]]$enrichr_hd = enrichr_results$ego

enrichr_results$plot +
  ggplot2::theme(axis.text.y = element_text(size = 8))

GSEA

We run a GSEA for all gene sets, from the full count matrix :

ranked_gene_list = aquarius::run_foldchange(Seurat::GetAssayData(subsobj, assay = "RNA", slot = "counts"),
                                            group1 = colnames(subsobj)[subsobj@active.ident %in% "HS"],
                                            group2 = colnames(subsobj)[subsobj@active.ident %in% "HD"])
names(ranked_gene_list) = gene_corresp$ID

gsea_results = aquarius::gsea_run(ranked_gene_list = ranked_gene_list,
                                  gene_sets = gene_sets[, c("gs_name", "ensembl_gene")],
                                  GSEA_p_val_thresh = 1)

list_results[[group_name]]$gsea = gsea_results

gsea_results@result %>%
  dplyr::filter(pvalue < 0.05) %>%
  dplyr::top_n(., n = 200, wt = abs(NES)) %>%
  dplyr::mutate(too_long = ifelse(nchar(ID) > 60, yes = TRUE, no = FALSE)) %>%
  dplyr::mutate(ID = stringr::str_sub(ID, end = 60)) %>%
  dplyr::mutate(ID = ifelse(too_long, yes = paste0(ID, "..."), no = ID)) %>%
  aquarius::gsea_plot(show_legend = TRUE) +
  ggplot2::labs(title = "GSEA using all genes (count matrix)") +
  ggplot2::theme(plot.title = element_text(size = 20))

We compute a fold change table to make a wordcloud :

fold_change = gene_corresp
colnames(fold_change) = c("gene_name", "ID")
rownames(fold_change) = fold_change$ID
fold_change$FC = ranked_gene_list[rownames(fold_change)]
fold_change$pct.1 = Seurat::GetAssayData(subsobj, assay = "RNA", slot = "counts")[, subsobj@active.ident == "HS"] %>%
  apply(., MARGIN = 1, FUN = function(one_row) {
    return( mean(one_row != 0) )
  })
fold_change$pct.2 = Seurat::GetAssayData(subsobj, assay = "RNA", slot = "counts")[, subsobj@active.ident == "HD"] %>%
  apply(., MARGIN = 1, FUN = function(one_row) {
    return( mean(one_row != 0) )
  })
fold_change$FC_x_pct = ifelse(fold_change$FC > 0,
                              yes = fold_change$FC * fold_change$pct.1,
                              no = fold_change$FC * fold_change$pct.2)

dim(fold_change) ; head(fold_change)

## [1] 16701     6

##                  gene_name              ID          FC        pct.1       pct.2
## ENSG00000238009 AL627309.1 ENSG00000238009 -0.79640583 0.0046728972 0.000000000
## ENSG00000237491 AL669831.5 ENSG00000237491  0.04968149 0.0919003115 0.044247788
## ENSG00000225880  LINC00115 ENSG00000225880 -1.01879825 0.0210280374 0.017699115
## ENSG00000230368     FAM41C ENSG00000230368  1.12415970 0.0389408100 0.004424779
## ENSG00000230699 AL645608.3 ENSG00000230699 -0.60376075 0.0054517134 0.000000000
## ENSG00000187634     SAMD11 ENSG00000187634 -3.60376075 0.0007788162 0.004424779
##                     FC_x_pct
## ENSG00000238009  0.000000000
## ENSG00000237491  0.004565744
## ENSG00000225880 -0.018031827
## ENSG00000230368  0.043775689
## ENSG00000230699  0.000000000
## ENSG00000187634 -0.015945844

We make the gsea plot for some gene sets :

gs_oi = c("REACTOME_EXPORT_OF_VIRAL_RIBONUCLEOPROTEINS_FROM_NUCLEUS",
          "GOBP_MICROTUBULE_ANCHORING",
          "GOBP_HISTONE_H3_K4_TRIMETHYLATION",
          "GOBP_HISTONE_H3_K4_MONOMETHYLATION",
          "GOBP_MITOPHAGY",
          "REACTOME_SEROTONIN_NEUROTRANSMITTER_RELEASE_CYCLE",
          "REACTOME_DOPAMINE_NEUROTRANSMITTER_RELEASE_CYCLE")

plot_list = make_gsea_plot(gsea_results, gs_oi, fold_change, "FC_x_pct")

patchwork::wrap_plots(plot_list, ncol = 2, widths = c(2,1.5))

ORS : HS vs HD

group_name = "ORS_HS_vs_HD"

We subset the Seurat object and change cell identities to sample type.

subsobj = subset(sobj, cluster_type == "ORS")
Seurat::Idents(subsobj) = subsobj$sample_type

table(subsobj$sample_type)

## 
##   HS   HD 
## 1563  459

DE

We identify DE genes between HS and HD :

mark = Seurat::FindMarkers(subsobj, ident.1 = "HS", ident.2 = "HD")

mark = mark %>%
  dplyr::filter(p_val_adj < 0.05) %>%
  dplyr::arrange(-avg_logFC, pct.1 - pct.2)

list_results[[group_name]]$mark = mark

dim(mark)

## [1] 145   5

head(mark, n = 20)

##                   p_val avg_logFC pct.1 pct.2     p_val_adj
## S100A7     1.177379e-34 1.3981213 0.390 0.083  1.966341e-30
## S100A9     8.042694e-80 1.3882495 0.831 0.346  1.343210e-75
## S100A8     6.069068e-63 1.2071418 0.784 0.353  1.013595e-58
## RPS26     8.289007e-157 1.0156704 0.971 0.950 1.384347e-152
## LGALS7B    1.751974e-26 0.9144582 0.636 0.503  2.925971e-22
## CCL2       4.001592e-35 0.8469394 0.592 0.279  6.683058e-31
## LTF        1.036558e-17 0.7369496 0.183 0.022  1.731156e-13
## MTRNR2L8   2.332458e-41 0.7044528 0.909 0.904  3.895438e-37
## IFI27      4.002962e-19 0.6856359 0.194 0.024  6.685347e-15
## TIMP1      1.883812e-25 0.6120579 0.566 0.342  3.146155e-21
## IFITM3     2.955422e-49 0.5618141 0.889 0.819  4.935850e-45
## CD74       9.221913e-27 0.5594045 0.347 0.096  1.540152e-22
## HLA-C      1.036948e-24 0.5483711 0.731 0.547  1.731807e-20
## MIF        1.151305e-28 0.5462406 0.403 0.159  1.922795e-24
## AKR1B10    7.040926e-44 0.5197609 0.372 0.033  1.175905e-39
## MTRNR2L12  1.018482e-25 0.4788943 0.951 0.943  1.700967e-21
## MTRNR2L10  1.055574e-11 0.4753293 0.589 0.538  1.762914e-07
## CXCL14     2.007187e-14 0.4495196 0.882 0.802  3.352203e-10
## LGALS7     5.171905e-12 0.4358938 0.232 0.096  8.637599e-08
## IL1R2      9.715791e-13 0.4197815 0.226 0.085  1.622634e-08

Up in HS

We explore enrichment in gene sets for HS population.

genes_of_interest = mark %>%
  dplyr::filter(avg_logFC > 0) %>%
  rownames()

enrichr_results = aquarius::run_enrichr(gene_names = genes_of_interest,
                                        gene_corresp = gene_corresp,
                                        gene_sets = gene_sets[, c("gs_name", "ensembl_gene")],
                                        make_plot = TRUE,
                                        plot_title = "Up-regulated in HS compared to HD")

list_results[[group_name]]$enrichr_hs = enrichr_results$ego

enrichr_results$plot +
  ggplot2::theme(axis.text.y = element_text(size = 8))

Up in HD

We explore enrichment in gene sets for HD population.

genes_of_interest = mark %>%
  dplyr::filter(avg_logFC < 0) %>%
  rownames()

enrichr_results = aquarius::run_enrichr(gene_names = genes_of_interest,
                                        gene_corresp = gene_corresp,
                                        gene_sets = gene_sets[, c("gs_name", "ensembl_gene")],
                                        make_plot = TRUE,
                                        plot_title = "Down-regulated in HS compared to HD")

list_results[[group_name]]$enrichr_hd = enrichr_results$ego

enrichr_results$plot +
  ggplot2::theme(axis.text.y = element_text(size = 8))

GSEA

We run a GSEA for all gene sets, from the full count matrix :

ranked_gene_list = aquarius::run_foldchange(Seurat::GetAssayData(subsobj, assay = "RNA", slot = "counts"),
                                            group1 = colnames(subsobj)[subsobj@active.ident %in% "HS"],
                                            group2 = colnames(subsobj)[subsobj@active.ident %in% "HD"])
names(ranked_gene_list) = gene_corresp$ID

gsea_results = aquarius::gsea_run(ranked_gene_list = ranked_gene_list,
                                  gene_sets = gene_sets[, c("gs_name", "ensembl_gene")],
                                  GSEA_p_val_thresh = 1)

list_results[[group_name]]$gsea = gsea_results

gsea_results@result %>%
  dplyr::filter(pvalue < 0.05) %>%
  dplyr::top_n(., n = 200, wt = abs(NES)) %>%
  dplyr::mutate(too_long = ifelse(nchar(ID) > 60, yes = TRUE, no = FALSE)) %>%
  dplyr::mutate(ID = stringr::str_sub(ID, end = 60)) %>%
  dplyr::mutate(ID = ifelse(too_long, yes = paste0(ID, "..."), no = ID)) %>%
  aquarius::gsea_plot(show_legend = TRUE) +
  ggplot2::labs(title = "GSEA using all genes (count matrix)") +
  ggplot2::theme(plot.title = element_text(size = 20))

We compute a fold change table to make a wordcloud :

fold_change = gene_corresp
colnames(fold_change) = c("gene_name", "ID")
rownames(fold_change) = fold_change$ID
fold_change$FC = ranked_gene_list[rownames(fold_change)]
fold_change$pct.1 = Seurat::GetAssayData(subsobj, assay = "RNA", slot = "counts")[, subsobj@active.ident == "HS"] %>%
  apply(., MARGIN = 1, FUN = function(one_row) {
    return( mean(one_row != 0) )
  })
fold_change$pct.2 = Seurat::GetAssayData(subsobj, assay = "RNA", slot = "counts")[, subsobj@active.ident == "HD"] %>%
  apply(., MARGIN = 1, FUN = function(one_row) {
    return( mean(one_row != 0) )
  })
fold_change$FC_x_pct = ifelse(fold_change$FC > 0,
                              yes = fold_change$FC * fold_change$pct.1,
                              no = fold_change$FC * fold_change$pct.2)

dim(fold_change) ; head(fold_change)

## [1] 16701     6

##                  gene_name              ID         FC       pct.1       pct.2
## ENSG00000238009 AL627309.1 ENSG00000238009 -0.2330146 0.005118362 0.004357298
## ENSG00000237491 AL669831.5 ENSG00000237491  0.7769695 0.079334613 0.043572985
## ENSG00000225880  LINC00115 ENSG00000225880  0.3519479 0.038387716 0.028322440
## ENSG00000230368     FAM41C ENSG00000230368  0.1550558 0.062699936 0.054466231
## ENSG00000230699 AL645608.3 ENSG00000230699 -0.8179771 0.005758157 0.008714597
## ENSG00000187634     SAMD11 ENSG00000187634 -1.3325503 0.003198976 0.008714597
##                     FC_x_pct
## ENSG00000238009 -0.001015314
## ENSG00000237491  0.061640571
## ENSG00000225880  0.013510475
## ENSG00000230368  0.009721990
## ENSG00000230699 -0.007128341
## ENSG00000187634 -0.011612639

We make the gsea plot for some gene sets :

gs_oi = c("GOBP_RESPONSE_TO_MOLECULE_OF_BACTERIAL_ORIGIN",
          "GOBP_HUMORAL_IMMUNE_RESPONSE",
          "HALLMARK_APOPTOSIS",
          "GOBP_GRANULOCYTE_CHEMOTAXIS",
          "GOBP_GRANULOCYTE_MIGRATION",
          "HALLMARK_INTERFERON_ALPHA_RESPONSE",
          "HALLMARK_INTERFERON_GAMMA_RESPONSE",
          "HALLMARK_TNFA_SIGNALING_VIA_NFKB",
          "HALLMARK_HYPOXIA",
          "GOBP_RESPONSE_TO_TRANSFORMING_GROWTH_FACTOR_BETA")

plot_list = make_gsea_plot(gsea_results, gs_oi, fold_change, "FC_x_pct")

patchwork::wrap_plots(plot_list, ncol = 2, widths = c(2,1.5))

Global heatmap

We represent all differentially expressed genes on a heatmap. First, we extract all DE genes :

features_oi = c(mark_to_label$gene_name,
                rownames(list_results$IBL_HS_vs_HD$mark),
                rownames(list_results$ORS_HS_vs_HD$mark)) %>%
  unique()

length(features_oi)

## [1] 259

We prepare the scaled expression matrix :

mat_expression = Seurat::GetAssayData(sobj, assay = "RNA", slot = "data")[features_oi, ]
mat_expression = Matrix::t(mat_expression)
mat_expression = dynutils::scale_quantile(mat_expression) # between 0 and 1
mat_expression = Matrix::t(mat_expression)
mat_expression = as.matrix(mat_expression) # not sparse

dim(mat_expression)

## [1]  259 3532

We prepare the heatmap annotation :

ha_top = ComplexHeatmap::HeatmapAnnotation(
  cluster_type = sobj$cluster_type,
  sample_type = sobj$sample_type,
  cluster = sobj$seurat_clusters,
  col = list(cluster_type = color_markers,
             sample_type = setNames(nm = c("HS", "HD"),
                                    c("#C55F40", "#2C78E6")),
             cluster = setNames(nm = levels(sobj$seurat_clusters),
                                aquarius::gg_color_hue(length(levels(sobj$seurat_clusters))))))

And the heatmap :

sobj$cell_group = paste0(sobj$cluster_type, sobj$sample_type) %>%
  as.factor()

ht = ComplexHeatmap::Heatmap(mat_expression,
                             col = aquarius::color_cnv,
                             # Annotation
                             top_annotation = ha_top,
                             # Grouping
                             column_order = sobj@meta.data %>%
                               dplyr::arrange(cluster_type, sample_type, seurat_clusters) %>%
                               rownames(),
                             column_split = sobj$cell_group,
                             column_gap = unit(c(0.01, 2, 0.01), "mm"),
                             column_title = NULL,
                             cluster_rows = TRUE,
                             cluster_columns = FALSE,
                             show_column_names = FALSE,
                             # Visual aspect
                             show_heatmap_legend = TRUE,
                             border = TRUE)

ComplexHeatmap::draw(ht,
                     merge_legend = TRUE,
                     heatmap_legend_side = "bottom",
                     annotation_legend_side = "bottom")

Save

We save the list of results :

saveRDS(list_results, file = paste0(out_dir, "/", save_name, "_list_results.rds"))

R Session

show

## R version 3.6.3 (2020-02-29)
## Platform: x86_64-pc-linux-gnu (64-bit)
## Running under: Ubuntu 20.04.6 LTS
## 
## Matrix products: default
## BLAS:   /usr/local/lib/R/lib/libRblas.so
## LAPACK: /usr/local/lib/R/lib/libRlapack.so
## 
## locale:
## [1] C
## 
## attached base packages:
##  [1] parallel  stats4    grid      stats     graphics  grDevices utils    
##  [8] datasets  methods   base     
## 
## other attached packages:
##  [1] org.Mm.eg.db_3.10.0   AnnotationDbi_1.48.0  IRanges_2.20.2       
##  [4] S4Vectors_0.24.4      Biobase_2.46.0        BiocGenerics_0.32.0  
##  [7] ComplexHeatmap_2.14.0 ggplot2_3.3.5         patchwork_1.1.2      
## [10] dplyr_1.0.7          
## 
## loaded via a namespace (and not attached):
##   [1] softImpute_1.4              graphlayouts_0.7.0         
##   [3] pbapply_1.4-2               lattice_0.20-41            
##   [5] haven_2.3.1                 vctrs_0.3.8                
##   [7] usethis_2.0.1               dynwrap_1.2.1              
##   [9] blob_1.2.1                  survival_3.2-13            
##  [11] prodlim_2019.11.13          dynutils_1.0.5             
##  [13] later_1.3.0                 DBI_1.1.1                  
##  [15] R.utils_2.11.0              SingleCellExperiment_1.8.0 
##  [17] rappdirs_0.3.3              uwot_0.1.8                 
##  [19] dqrng_0.2.1                 jpeg_0.1-8.1               
##  [21] zlibbioc_1.32.0             pspline_1.0-18             
##  [23] pcaMethods_1.78.0           mvtnorm_1.1-1              
##  [25] htmlwidgets_1.5.4           GlobalOptions_0.1.2        
##  [27] future_1.22.1               UpSetR_1.4.0               
##  [29] laeken_0.5.2                leiden_0.3.3               
##  [31] clustree_0.4.3              scater_1.14.6              
##  [33] irlba_2.3.3                 markdown_1.1               
##  [35] DEoptimR_1.0-9              tidygraph_1.1.2            
##  [37] Rcpp_1.0.9                  readr_2.0.2                
##  [39] KernSmooth_2.23-17          carrier_0.1.0              
##  [41] promises_1.1.0              gdata_2.18.0               
##  [43] DelayedArray_0.12.3         limma_3.42.2               
##  [45] graph_1.64.0                RcppParallel_5.1.4         
##  [47] Hmisc_4.4-0                 fs_1.5.2                   
##  [49] RSpectra_0.16-0             fastmatch_1.1-0            
##  [51] ranger_0.12.1               digest_0.6.25              
##  [53] png_0.1-7                   sctransform_0.2.1          
##  [55] cowplot_1.0.0               DOSE_3.12.0                
##  [57] here_1.0.1                  TInGa_0.0.0.9000           
##  [59] ggraph_2.0.3                pkgconfig_2.0.3            
##  [61] GO.db_3.10.0                DelayedMatrixStats_1.8.0   
##  [63] gower_0.2.1                 ggbeeswarm_0.6.0           
##  [65] iterators_1.0.12            DropletUtils_1.6.1         
##  [67] reticulate_1.26             clusterProfiler_3.14.3     
##  [69] SummarizedExperiment_1.16.1 circlize_0.4.15            
##  [71] beeswarm_0.4.0              GetoptLong_1.0.5           
##  [73] xfun_0.35                   bslib_0.3.1                
##  [75] zoo_1.8-10                  tidyselect_1.1.0           
##  [77] reshape2_1.4.4              purrr_0.3.4                
##  [79] ica_1.0-2                   pcaPP_1.9-73               
##  [81] viridisLite_0.3.0           rtracklayer_1.46.0         
##  [83] rlang_1.0.2                 hexbin_1.28.1              
##  [85] jquerylib_0.1.4             dyneval_0.9.9              
##  [87] glue_1.4.2                  RColorBrewer_1.1-2         
##  [89] matrixStats_0.56.0          stringr_1.4.0              
##  [91] lava_1.6.7                  europepmc_0.3              
##  [93] DESeq2_1.26.0               recipes_0.1.17             
##  [95] labeling_0.3                httpuv_1.5.2               
##  [97] class_7.3-17                BiocNeighbors_1.4.2        
##  [99] DO.db_2.9                   annotate_1.64.0            
## [101] jsonlite_1.7.2              XVector_0.26.0             
## [103] bit_4.0.4                   mime_0.9                   
## [105] aquarius_0.1.5              Rsamtools_2.2.3            
## [107] gridExtra_2.3               gplots_3.0.3               
## [109] stringi_1.4.6               processx_3.5.2             
## [111] gsl_2.1-6                   bitops_1.0-6               
## [113] cli_3.0.1                   batchelor_1.2.4            
## [115] RSQLite_2.2.0               randomForest_4.6-14        
## [117] tidyr_1.1.4                 data.table_1.14.2          
## [119] rstudioapi_0.13             GenomicAlignments_1.22.1   
## [121] nlme_3.1-147                qvalue_2.18.0              
## [123] scran_1.14.6                locfit_1.5-9.4             
## [125] scDblFinder_1.1.8           listenv_0.8.0              
## [127] ggthemes_4.2.4              gridGraphics_0.5-0         
## [129] R.oo_1.24.0                 dbplyr_1.4.4               
## [131] TTR_0.24.2                  readxl_1.3.1               
## [133] lifecycle_1.0.1             timeDate_3043.102          
## [135] ggpattern_0.3.1             munsell_0.5.0              
## [137] cellranger_1.1.0            R.methodsS3_1.8.1          
## [139] proxyC_0.1.5                visNetwork_2.0.9           
## [141] caTools_1.18.0              codetools_0.2-16           
## [143] ggwordcloud_0.5.0           GenomeInfoDb_1.22.1        
## [145] vipor_0.4.5                 lmtest_0.9-38              
## [147] msigdbr_7.5.1               htmlTable_1.13.3           
## [149] triebeard_0.3.0             lsei_1.2-0                 
## [151] xtable_1.8-4                ROCR_1.0-7                 
## [153] BiocManager_1.30.10         scatterplot3d_0.3-41       
## [155] abind_1.4-5                 farver_2.0.3               
## [157] parallelly_1.28.1           RANN_2.6.1                 
## [159] askpass_1.1                 GenomicRanges_1.38.0       
## [161] RcppAnnoy_0.0.16            tibble_3.1.5               
## [163] ggdendro_0.1-20             cluster_2.1.0              
## [165] future.apply_1.5.0          Seurat_3.1.5               
## [167] dendextend_1.15.1           Matrix_1.3-2               
## [169] ellipsis_0.3.2              prettyunits_1.1.1          
## [171] lubridate_1.7.9             ggridges_0.5.2             
## [173] igraph_1.2.5                RcppEigen_0.3.3.7.0        
## [175] fgsea_1.12.0                remotes_2.4.2              
## [177] scBFA_1.0.0                 destiny_3.0.1              
## [179] VIM_6.1.1                   testthat_3.1.0             
## [181] htmltools_0.5.2             BiocFileCache_1.10.2       
## [183] yaml_2.2.1                  utf8_1.1.4                 
## [185] plotly_4.9.2.1              XML_3.99-0.3               
## [187] ModelMetrics_1.2.2.2        e1071_1.7-3                
## [189] foreign_0.8-76              withr_2.5.0                
## [191] fitdistrplus_1.0-14         BiocParallel_1.20.1        
## [193] xgboost_1.4.1.1             bit64_4.0.5                
## [195] foreach_1.5.0               robustbase_0.93-9          
## [197] Biostrings_2.54.0           GOSemSim_2.13.1            
## [199] rsvd_1.0.3                  memoise_2.0.0              
## [201] evaluate_0.18               forcats_0.5.0              
## [203] rio_0.5.16                  geneplotter_1.64.0         
## [205] tzdb_0.1.2                  caret_6.0-86               
## [207] ps_1.6.0                    DiagrammeR_1.0.6.1         
## [209] curl_4.3                    fdrtool_1.2.15             
## [211] fansi_0.4.1                 highr_0.8                  
## [213] urltools_1.7.3              xts_0.12.1                 
## [215] GSEABase_1.48.0             acepack_1.4.1              
## [217] edgeR_3.28.1                checkmate_2.0.0            
## [219] scds_1.2.0                  cachem_1.0.6               
## [221] npsurv_0.4-0                babelgene_22.3             
## [223] rjson_0.2.20                openxlsx_4.1.5             
## [225] ggrepel_0.9.1               clue_0.3-60                
## [227] rprojroot_2.0.2             stabledist_0.7-1           
## [229] tools_3.6.3                 sass_0.4.0                 
## [231] nichenetr_1.1.1             magrittr_2.0.1             
## [233] RCurl_1.98-1.2              proxy_0.4-24               
## [235] car_3.0-11                  ape_5.3                    
## [237] ggplotify_0.0.5             xml2_1.3.2                 
## [239] httr_1.4.2                  assertthat_0.2.1           
## [241] rmarkdown_2.18              boot_1.3-25                
## [243] globals_0.14.0              R6_2.4.1                   
## [245] Rhdf5lib_1.8.0              nnet_7.3-14                
## [247] RcppHNSW_0.2.0              progress_1.2.2             
## [249] genefilter_1.68.0           statmod_1.4.34             
## [251] gtools_3.8.2                shape_1.4.6                
## [253] HDF5Array_1.14.4            BiocSingular_1.2.2         
## [255] rhdf5_2.30.1                splines_3.6.3              
## [257] AUCell_1.8.0                carData_3.0-4              
## [259] colorspace_1.4-1            generics_0.1.0             
## [261] base64enc_0.1-3             dynfeature_1.0.0           
## [263] smoother_1.1                gridtext_0.1.1             
## [265] pillar_1.6.3                tweenr_1.0.1               
## [267] sp_1.4-1                    ggplot.multistats_1.0.0    
## [269] rvcheck_0.1.8               GenomeInfoDbData_1.2.2     
## [271] plyr_1.8.6                  gtable_0.3.0               
## [273] zip_2.2.0                   knitr_1.41                 
## [275] latticeExtra_0.6-29         biomaRt_2.42.1             
## [277] fastmap_1.1.0               ADGofTest_0.3              
## [279] copula_1.0-0                doParallel_1.0.15          
## [281] vcd_1.4-8                   babelwhale_1.0.1           
## [283] openssl_1.4.1               scales_1.1.1               
## [285] backports_1.2.1             ipred_0.9-12               
## [287] enrichplot_1.6.1            hms_1.1.1                  
## [289] ggforce_0.3.1               Rtsne_0.15                 
## [291] shiny_1.7.1                 numDeriv_2016.8-1.1        
## [293] polyclip_1.10-0             lazyeval_0.2.2             
## [295] Formula_1.2-3               tsne_0.1-3                 
## [297] crayon_1.3.4                MASS_7.3-54                
## [299] pROC_1.16.2                 viridis_0.5.1              
## [301] dynparam_1.0.0              rpart_4.1-15               
## [303] zinbwave_1.8.0              compiler_3.6.3             
## [305] ggtext_0.1.0